Harnessing data and analytics for citizen health and wellbeing

The human services sector has rich datasets that can help caseworkers appreciate the full picture of concern. Now it needs to embrace safe data sharing.

Given the human services’ progress and the will to collect, share and use data more effectively, the sector is set on a good course to realise its ambition to holistically meet the needs of our most vulnerable citizens by working collaboratively across the ecosystem.

During COVID-19, Australia’s human services agencies have been pushed to rapidly adopt digital technology, with 59% of respondents of a survey of Australia’s human services professionals saying their agency’s use of digital technologies and data solutions has increased.

Even before the pandemic, government agencies and providers recognised that reliable, cross-agency data is critical to help make better decisions and design more informed policy options. COVID-19 magnified that need, catalysing great efforts to make easy-to-visualise data accessible to decision-makers at all levels.

Importantly, citizens’ trust in government increased this year, when they saw their elected representatives harness real-time data and step in swiftly to create a safety net for those affected by the health and economic shocks of COVID-19.

Digital and data solutions are not temporary

That said, EY global research into human services agencies found 46% of Australian respondents saying that “introducing digital technologies and data solutions was a temporary measure to help our organisation get through the pandemic period”.

It’s a confounding finding when respondents were largely positive about the way agencies invested rapidly in technology upgrades to support virtual contact with vulnerable individuals and families. Around two-thirds of Australian respondents said the use of digital technologies and data solutions had improved access to care, delivered better outcomes for customers and made staff more productive since the outbreak of COVID-19.

One of the biggest issues faced during 2020 was the risk that many vulnerable citizens became invisible due to COVID-19 restrictions. A major concern for human services agencies throughout the pandemic was how to reach customers who were previously engaged through face-to-face contact, such as home or office appointments, or reporting through school or hospital referrals. In particular, the pandemic brought a whole new cohort of customers into the risk of vulnerability due to loss of jobs, homes or loved ones, and agencies weren’t able to detect them in their usual way.

There remains great concern that the most vulnerable and disadvantaged, including those without stable internet access, will ‘fall through the cracks’ if services become increasingly digital. A myth is emerging that the act of turning a paper file into binary code is somehow disadvantageous to the work of human services.

Nothing could be further from the truth.

Of course, human services agencies will need to continue face-to-face contact for the most at-risk cohorts. For some individuals, families and communities, face-to-face will remain the best medium for engagement. No-one is suggesting home and office visits should cease.

But customers now expect to access help online and by phone, and indeed, some prefer this. Some agencies are finding they can improve input to caseworker assessments and decision-making through increased use of digitising, analysing and sharing data — building a richer picture of need and customer preference as they protect and help vulnerable citizens.

Still in the development stages

Currently, despite all the progress being made across government, many human services caseworkers do not have access to all the key data held by their own agency or adjacent agencies within the same government tier.

At the extreme, we have examples of agency collaboration where colleagues from different agencies (like child protection, health, police and justice) sit in co-located offices logged into different, unconnected systems. Their usual means of sharing data is talking to the cross-agency colleagues sitting next to them.

Most human services agencies do share data electronically, but there are some still reliant on phone or fax. As a result, frustrated customers still have to repeat their story to different agencies — and caseworkers frequently visit people without knowing what happened with another agency the week before.

Every public sector agent who intersects with the same citizen should have access to all the appropriate data relevant to that citizen in real time.

Leading the way, New Zealand has a shared dataset across its entire population that is used by all government agencies to shape policy and frontline services. As just one example, Oranga Tamariki (the NZ child protection agency) uses the data to:

• understand vulnerable families and children
• design its service offer to target services at the most vulnerable
• intervene as early as possible in those families’ lives before crisis occurs.

Citizens are not supported

Perhaps even more concerning, until we crack data sharing, we cannot use the most advanced digital tools to support human services goals.

If consumers can find and book accommodation on Airbnb in under a minute, why are our caseworkers spending hours calling around to find placements for at-risk children? If a smart crop system can predict when a farmer needs to order more nitrogen, why can’t we actively monitor key indicators like school absenteeism or hospital visits to predict when children are likely to become at risk?

The technology exists to do all of this — and so much more. And we don’t even need to gather more information.

Agencies can and do integrate existing datasets to understand cohorts and interactions across agencies and communities, and design better programs, services and responses. By combining multiple datasets from across systems and reviewing against historic patterns, new tools can uncover early markers that signal a child or family is at risk.

In the UK, by running predictive analytics over publicly available data, one London borough can now identify six to eight weeks in advance that a family may become homeless — and nine months ahead of time that a child may need to be taken into care.

In Brazil, vulnerable families are identified before they become homeless by analysing welfare payments and changes in food bank usage patterns.

Wherever data sharing is prevalent, agencies are able to respond to early warning signals and offer support before families reach crisis point. Such interventions not only improve citizen wellbeing — and sometimes save lives — they are also substantially more cost-effective than waiting until a crisis occurs.

Automated data collecting, sharing and visualisation tools could save caseworkers 30 to 50% of their administrative time in some instances, freeing them up to spend more time with at-risk children and families.

In New Zealand, the Family Harm app has replaced a 13-page paper form that used to take police officers 40 minutes to complete. Information that used to take days to reach police records is now uploaded instantly — often during an interaction. The app also gives frontline officers useful information such as historical episodes at the same location. Police describe having this additional context as an “eyes wide open” approach that is paying huge dividends.

Caseworkers need more robust data

Many government agencies have formed cross-agency intelligence functions and integrated datasets to enable real-time analysis of risk and forecasting of need — both to inform policy and service system design and also enable rapid decision-making during operational service delivery.

At the same time, the Commonwealth has continued to invest in building data capability across the Australian Public Service through initiatives such as the Multi-Agency Data Integration Project and the data professional stream strategy.

But nowhere is this more important than in human services, which relies heavily on its caseworkers exercising judgement. Their job is to make the best decision they can in any given situation. And, despite decades of experience, enormous dedication and immensely good intentions, caseworkers are only human. For a multitude of reasons, some decisions are made without the benefit of all the key information relevant to the case.

These indispensable frontline staff do an amazing job, often burdened by heavy workloads. They don’t always have a lot of time to triage a new reported risk of harm or make a decision on a family in their caseload. Also, because at-risk families are seen by many agencies, even when caseworkers see risk markers they may assume someone else is already dealing with the issue — a variation of the well-documented ‘bystander effect’.

Five myths holding back data sharing

In both our formal research interviewing frontline workers and end users, and our work with clients, we continually hear people speak of a desire to move to holistic, integrated service delivery. And the pandemic has demonstrated that people can react quickly to new technologies, with 58% of Australia’s human services respondents thinking their agency staff quickly adapted to the use of new digital technology and data solutions.

We believe if people had time to stand back and think about what’s possible and understand the reality of data sharing, resistance would fade away.

To this point, let us dispel once and for all the myths and concerns around data sharing. Anecdotal evidence suggests some human services frontline workers are reluctant to share data because they believe:

1. Privacy issues have not been solved. Most states have already passed legislation that enables information sharing. In contrast to historical precedent, privacy commissioners are now encouraging data sharing with the user’s consent and where it makes sense. Caseworkers should also be aware that a lot of the data shared is anonymised. Datasets can be linked in a central way that protects the identity of individuals but brings together critical data from different systems across agencies. This can then be presented back to case workers in automated case notes, putting together all the touchpoints an individual or household has with human services agencies.
2. People can’t be reduced to numbers. This is true, but data sharing isn’t just about the numbers. Written case notes can be digitised and searched, giving caseworkers a summary of decades of notes in seconds.
3. They need to see for themselves. Some caseworkers have learned from experience not to trust anything they haven’t seen with their own eyes. Without downplaying the importance of direct involvement in assessment, any additional information is an important resource, even if it comes with a question mark. Over time, as cross-agency collaboration and data sharing become more widespread, trust will build and caseworkers will become more confident about the information the system delivers to their phone before they walk into a home.
4. Digital is less secure than paper. Many people who are resisting digitising and sharing data have no qualms about handing over a paper file or faxing information to another agency. Arguably, properly protected digital records are at least as or even more secure than their paper equivalents.
5. Data and automation will replace the human touch. People fear that automation is the precursor to removing human judgment from human services. Yet the likelihood is that the opposite is true. In the mining industry, the introduction of automation has actually created jobs. The future of human services will continue to rely on human judgement and the need for a human touch. The technology being proposed will simply make it easier for people do their jobs well and give them rapid access to useful information that will make them even more effective.

How to keep digital momentum going

Holistic, integrated service delivery enabled by data sharing is the key to unlocking the potential in what the human services sector already has in place. To embrace the power of data to protect our most vulnerable citizens, agencies and departments should:

• set the vision from the top and embed process and practice signals that encourage data sharing
• open up the system to more flexible funding arrangements and new models of accountability to remove blockers between agencies and providers
• agree on new, shared measurements of success against which every agency and provider is held accountable
• create a culture that rewards individuals when they make a choice to try something new and think beyond traditional ways of planning and executing programs
• make investment in digital literacy as important as that in other professional development.
   

Ultimately, all these changes need to be driven by a desire to focus outcomes on an end user’s real needs and wellbeing. It will then be that individual who determines what those outcomes should be. As one end user said, wouldn’t it be fantastic if all organisations could “get on the same page and work for us”. That will not happen until human services embraces data sharing at every level.

Let’s not wait for the next crisis to drive change forward. Human services has everything it needs to build on the digital ‘muscle’ it has built over the last nine months and start to understand and make brilliant use of the rich data it holds.

EY Partner Dr Sonia Sharp holds a Ph.D. in educational psychology and has worked in senior government positions in Australia and the UK.

Survey methodology

The survey was conducted using an online interview administered to members of the YouGov Plc UK panel of 800,000+ individuals. Fieldwork was undertaken between 3–29 September 2020 with 2313 health and human services managers spread across Australia, India, Italy, New Zealand, UAE, US and the US. In Australia, the cohort included 69 human services managers.

Image credit: ©stock.adobe.com/au/flashmovie